{"id":"85636774-6154-41b5-92cb-6e15abc77c6b","arxiv_id":"2501.13634","paper_version":1,"verdict":"CONDITIONAL","confidence":"MODERATE","novelty_score":4.0,"correctness_risk":"medium","formal_verification":"none","parameter_count":5,"one_line_summary":"This paper reports MBE growth and 300 K Hall characterization of 12.5-500 nm InAs films on InAlAs metamorphic buffers, and uses a two-channel model to attribute most thin-film charge to surface and interface layers.","lead":"Researchers grew very thin InAs layers, from 12.5 to 500 nanometers, on insulating InAlAs metamorphic buffers and measured how crystal quality and electrical transport depend on thickness. A two-layer conduction model splits the charge into surface, bulk, and interface contributions, which matters for designing InAs-based superconducting quantum devices.","discovery_kind":"extension","skeptic_critique":{"model":"deepseek-v4-flash","headline":"The quantitative surface/interface decomposition is fit-dependent: Eq. (8) sets f=1 although the authors' own consistency check gives 1<f<2, and it assumes constant n_bs although the text says n_bs should vary with t; the quoted sheet densities are not uniquely determined.","rationale":"The reader's CONDITIONAL verdict already identifies the constant-n_bs/t_int assumption as the weak point. I agree with the direction but would sharpen it: the most direct load-bearing issue is the f=1 approximation. Equation (8) is obtained from Eq. (7) by setting f=1; the authors later compute 1<f<2 for all samples, so the fit used to extract n_int t_int ignores a factor that the paper itself shows is not unity. This is an internal consistency problem, not a disagreement with outside consensus. The constant-n_bs assumption compounds it: the text warns that n_bs should vary with t, yet the fit averages over that variation. The capping experiment is a good independent control for the surface channel, which is why I do not regard the qualitative conclusion as baseless. But the quoted numbers—especially n_int t_int=(4.8±0.3)×10^12 cm^-2—are conditional on exactly the two assumptions under test. A variable-field Hall/mobility-spectrum measurement would settle whether the two-channel decomposition with these densities actually describes the data. Since this is a request for stronger quantitative support rather than a demonstration of a wrong central result, the reader's CONDITIONAL verdict should stand.","tokens_in":13946,"tokens_out":12431,"duration_ms":109125,"concrete_test":"Perform variable-magnetic-field Hall measurements at 300 K on samples C, E, and one capped sample, and fit the field-dependent Hall coefficient and longitudinal conductivity with the two-layer Petritz model (Eqs. 6 and 9), treating f as a free parameter rather than setting it to 1. If the best-fit interface and surface sheet densities fall outside the quoted (4.8±0.3)×10^12 cm^-2 and (2.2±0.5)×10^12 cm^-2 by more than the stated errors, the f=1/constant-n_bs decomposition is not established; if they agree, the concern is resolved.","verdict_should_be":"UNCHANGED","load_bearing_attack":"The quantitative claim—n_s t_s = (2.2±0.5)×10^12 cm^-2, n_int t_int = (4.8±0.3)×10^12 cm^-2, and hence surface+interface dominance for t<100 nm—rests on the two-channel equations (6)–(9). Two uncontrolled approximations enter. First, Eq. (7) defines a factor f that is set to 1 in Eq. (8), but the authors' own consistency check after the fits returns 1<f<2 for every thickness. Equation (8) is exactly the f=1 case, so the fitted quantities n_bs and (n_int−n_bs)t_int are biased whenever f is not unity; a factor-of-~2 uncertainty in f propagates into the extracted interface sheet density. Second, the fit assumes n_bs is independent of t, although the paper explicitly says 'we do not expect n_bs to be constant as a function of t' because for t≲L_D the bulk–surface layer is surface-dominated. The resulting n_bs is only an average, so the data do not identify a constant bulk channel plus a constant interface channel. The acknowledged failure of the model for samples A and B (12.5 nm and 25 nm) occurs exactly in the thickness range where the surface/interface-dominance claim should be tested. The mobility curve in Fig. 8 is computed from the same best-fit parameters and therefore is not an independent check. The capping experiment (samples H/I) is a good control for the surface contribution, but it does not constrain the interface-layer parameters.","agreement_with_reader":"partial"},"referee_report":{"model":"deepseek-v4-flash","summary":"The paper reports the MBE growth of thin InAs layers (12.5–500 nm) on InxAl1−xAs metamorphic buffers and characterizes them with high-resolution X-ray diffraction and room-temperature van der Pauw transport measurements. The structural part shows that perpendicular strain and XRD FWHM scale roughly as 1/t. The transport part proposes a two-parallel-channel model with a constant interface layer and a bulk–surface layer. From fits to Equations (8) and (9), the authors extract a surface sheet density n_s t_s = (2.2 ± 0.5) × 10^12 cm^-2 and an interface sheet density n_int t_int = (4.8 ± 0.3) × 10^12 cm^-2, and conclude that for t ≲ 100 nm most of the charge in the InAs film resides in the surface and interface regions. The capping experiment (samples H and I) is used as a control for the surface contribution.","tokens_in":14324,"tokens_out":2473,"duration_ms":22677,"significance":"The structural characterization is solid: the 1/t dependence of strain and FWHM is supported by the data and the comparison with InAs grown directly on GaAs is informative. The capping experiment is a well-designed control and provides a direct estimate of the surface sheet density. If the quantitative transport decomposition were reliable, the extracted surface and interface sheet densities would be useful for designing InAs-based devices on insulating buffers. However, the decomposition rests on several approximations, some acknowledged in the text, that are not quantitatively propagated. The central quantitative claim—surface-plus-interface dominance for thin films with the stated sheet densities—is therefore not yet established to the standard of the structural results.","major_comments":[{"comment":"Setting the factor f to 1 in Equation (8) is not justified by the authors' own consistency check, which reports 1 < f < 2 for every thickness. Because f multiplies the expression for n, the fitted parameters n_bs and (n_int − n_bs)·t_int are biased whenever f differs from unity. The authors should either include f in the fitting procedure or propagate the uncertainty in f into the extracted sheet densities; otherwise the quoted n_int·t_int = (4.8 ± 0.3) × 10^12 cm^-2 does not reflect the true uncertainty of the model.","section":"Section 3.2, Equations (7)–(8)"},{"comment":"The model fitting assumes a constant n_bs, yet the text explicitly states that n_bs is not expected to be constant as a function of t because for t ≲ L_D the bulk–surface layer is surface-dominated. The fitted n_bs is therefore only a thickness-averaged value, and the data do not identify a constant bulk–surface channel plus a constant interface channel. This weakens the interpretation of the extracted interface sheet density and the conclusion that most of the charge is in the surface and interface layers for t < 100 nm.","section":"Section 3.2, text after Equation (8)"},{"comment":"The solid line in Figure 8 is computed from Equation (10) using the best-fit parameters obtained from Equations (8) and (9), which are fits to the same 300 K dataset. The agreement is therefore a re-plot of fitted quantities, not an independent validation of the two-channel model. An out-of-sample test, such as excluding one thickness from the fit and predicting it, would be needed to assess predictive power.","section":"Figure 8 and Equation (10)"},{"comment":"The model is acknowledged to fail for the two thinnest samples because the interface layer is expected to be less charged near the critical thickness. These are precisely the thicknesses in the range where the claim that surface and interface layers dominate (t ≲ 100 nm) is most consequential. The limitation should be stated as a caveat on the main conclusion, and the quantitative sheet densities should not be presented as applying to the full thickness range without this caveat.","section":"Section 3.2, samples A and B (12.5 nm and 25 nm)"},{"comment":"The interface layer thickness t_int is assumed to be ~10 nm for all samples without direct measurement, and the extracted n_int ≈ 4.8 × 10^18 cm^-3 scales inversely with t_int. If t_int varies with strain relaxation or dislocation density—which changes strongly from sample A to G—the decomposition into surface and interface charge is not uniquely determined. The authors should provide a sensitivity analysis or an independent estimate of t_int.","section":"Section 3.2, interface layer thickness t_int"}],"minor_comments":[{"comment":"The phrase 'Differently than InAs substrates' is non-standard English; consider 'Unlike InAs substrates'.","section":"Abstract"},{"comment":"The square brackets in Equation (10) are typeset inconsistently (the numerator uses an opening square bracket and the denominator uses a ceiling-like symbol). Please check the mathematical notation.","section":"Section 3.2, Equation (10)"},{"comment":"The fit line in Figure 6 would be clearer with the fit parameters and their uncertainties printed in the caption, rather than only in the text.","section":"Figure 6"},{"comment":"The error on relaxation is not reported; adding it would help evaluate the strain relaxation trends.","section":"Table 1"}],"recommendation":"major_revision","confidential_remarks":"The structural results and the capping experiment are useful contributions. The main issue is that the quantitative transport decomposition is presented with a precision that the model's own acknowledged approximations do not support. The authors should be asked to refit with f included, propagate the t_int uncertainty, and clearly separate the independent structural conclusions from the model-dependent transport decomposition. This is fixable within the manuscript's scope, so major revision rather than rejection seems appropriate."},"author_rebuttal":null,"desk_editor":{"model":"deepseek-v4-flash","letter":"First, what's worth taking from this paper: a systematic MBE growth series of InAs on InAlAs metamorphic buffers, from 12.5 to 500 nm, with clean XRD showing 1/t strain relaxation and FWHM dominated by thickness broadening. The capping experiment (samples H/I) is a sensible control and gives an independent estimate of the surface sheet density, (2.2±0.5)e12 cm^-2. The authors are also candid about the model's limitations, which is more than many papers do.\n\nNow the soft spots. The quantitative decomposition into surface and interface channels rests on Eq. (8), which sets the factor f to 1. The authors' own consistency check after the fits gives 1<f<2, so the extracted interface sheet density (4.8±0.3)e12 cm^-2 carries an unquantified systematic error of roughly a factor of two. The fit also assumes n_bs is constant, which the text explicitly says is not expected; the fitted n_bs is just an average. And the model fails for the two thinnest samples (12.5 and 25 nm), exactly the range where the surface/interface dominance claim should be tested. For 50 and 100 nm the sum of surface+interface sheet densities does match the measured Ns, so the qualitative claim holds there, but the statement 'for t<100 nm most of the charge is due to surface and interface' is too broad. Figure 8 is not an independent check because the mobility curve is computed from the same best-fit parameters.\n\nNone of this is fatal. The growth and structural data are solid, the paper is honest, and the capping experiment is a genuine independent handle on the surface contribution. But the headline numbers should be treated as order-of-magnitude estimates, not quantitative results. A serious referee should ask for the f-factor to be folded into the uncertainty, or for direct measurement of t_int (e.g., TEM).\n\nWho is this for? Groups engineering InAs-based hybrid superconductor devices, who need to know where the carriers sit. They will get a useful dataset and a reasonable qualitative picture. I would cite it for the growth and structural results, not for the extracted sheet densities. Recommendation: yes, send it to peer review. It is a real experimental study with no red flags, and the limitations can be addressed in revision.","headline":"Solid growth and XRD, useful dataset, but the quantitative surface/interface decomposition is a self-consistent fit rather than an independent measurement; the extracted sheet densities carry an unquantified factor-of-two bias.","tokens_in":14914,"tokens_out":3320,"would_cite":true,"duration_ms":29736,"reading_group":"maybe","serious_thinker":"yes","would_accept_peer_review":true},"rs_alignment":null,"lean_confirmation":null,"pith_extraction":{"msc":[],"pacs":[],"model":"deepseek-v4-flash","headline":"Thin InAs films on metamorphic buffers conduct mainly through a defected interface layer and a surface accumulation layer, not through the bulk, for thicknesses below about 100 nm.","keywords":["InAs","InAlAs metamorphic buffer","molecular beam epitaxy","van der Pauw Hall","parallel-conduction model","surface charge accumulation","interface dislocations","superconductor-semiconductor devices"],"falsifier":"Perform low-temperature Hall and mobility-spectrum measurements on the 50 nm and 100 nm samples to resolve the two channels independently; if the resolved surface and interface sheet densities do not sum to the measured total sheet density, or if $n_{\\mathrm{bs}}$ extracted from films of different thickness is found to change systematically as strain relaxes, the fixed-$n_{\\mathrm{bs}}$ two-channel decomposition in Eq. (8) is not a unique description.","tokens_in":13729,"feed_emoji":"⚡","tokens_out":9606,"duration_ms":78391,"temperature":0.7,"pith_summary":"This paper tries to establish that the room-temperature transport of thin InAs layers grown by molecular beam epitaxy on InAlAs metamorphic buffers is controlled by two thin conducting regions—a surface accumulation layer and a defected interface layer—rather than by the bulk of the film. It reports growth of InAs films from 12.5 nm to 500 nm, structural characterization by high-resolution X-ray diffraction, and Hall measurements in the van der Pauw geometry at 300 K. A two-parallel-channel model reproduces the thickness dependence of carrier concentration and mobility, and capped-reference samples separate the surface contribution from the interface contribution. If the model is right, for films below about 100 nm the usable charge in a device sits almost entirely at the surface and interface, which matters for the design of gate-tunable superconductor-semiconductor quantum devices on insulating buffers.","feed_headline":"Most charge in thin InAs films sits at surface and interface","feed_subtitle":"Hall data on 12.5-500 nm layers pin surface and interface sheet densities near 7 x 10^12 cm^-2.","key_machinery":"The load-bearing object is the two-parallel-channel conduction model, a reduction of Petritz's two-layer Hall-coefficient formula to $n = n_{\\mathrm{bs}} + (n_{\\mathrm{int}} - n_{\\mathrm{bs}}) t_{\\mathrm{int}}/t$ by setting the correction factor $f$ to one. The model splits the film into a fixed ~10 nm defected interface layer, whose thickness is identified with the critical thickness of InAs on In$_{0.84}$Al$_{0.16}$As, and a 'bulk–surface' layer of thickness $t - t_{\\mathrm{int}}$ whose carrier concentration $n_{\\mathrm{bs}}$ is treated as a constant average over the dataset. The same decomposition is applied to conductivity, and the two fits combine through $\\mu_H = \\sigma/(ne)$ into a mobility-versus-thickness curve that tracks the measured Hall mobility. Capped samples with a 5 nm In$_{0.84}$Al$_{0.16}$As layer supply the independent surface-charge estimate that completes the decomposition.","core_discovery":"The central claim is that the thickness dependence of the Hall carrier concentration in InAs on InAlAs metamorphic buffers is described by two parallel conduction channels: a heavily defected interface layer about 10 nm thick with volume concentration $n_{\\mathrm{int}} \\simeq 4.8\\times10^{18}$ cm$^{-3}$, and a 'bulk–surface' layer with an average concentration $n_{\\mathrm{bs}} = (5.1 \\pm 1.4)\\times10^{16}$ cm$^{-3}$. The fit of $n = n_{\\mathrm{bs}} + (n_{\\mathrm{int}} - n_{\\mathrm{bs}}) t_{\\mathrm{int}}/t$ to samples from 12.5 to 500 nm gives an interface sheet density $n_{\\mathrm{int}}t_{\\mathrm{int}} = (4.8 \\pm 0.3)\\times10^{12}$ cm$^{-2}$, and capped samples give a surface sheet density $n_{\\mathrm{s}}t_{\\mathrm{s}} = (2.2 \\pm 0.5)\\times10^{12}$ cm$^{-2}$. The two sum to $(7.0 \\pm 0.6)\\times10^{12}$ cm$^{-2}$, matching the measured sheet densities of the 50 nm and 100 nm films, which supports the conclusion that for $t \\lesssim 100$ nm most of the charge lies at the surface and interface rather than in the bulk.","pith_inferences":["If the same two channels persist at cryogenic temperatures, gate-tunable superconductor-semiconductor devices on these buffers would start from a fixed surface/interface electron budget of roughly $7\\times10^{12}$ cm$^{-2}$, which sets the scale for depletion-gate design.","A direct test of the dislocation picture would be to correlate the fitted interface sheet density with threading-dislocation densities measured by X-ray topography or cross-sectional TEM on identically grown films; the model implies a monotonic relationship.","Because $n_{\\mathrm{bs}}$ is fitted as a thickness-independent average while strain relaxation clearly evolves with thickness, the quoted surface and interface sheet densities are best read as effective parameters of a two-layer model, not as uniquely determined physical constants."],"forward_implications":["For 50 nm and 100 nm films, the measured sheet carrier concentrations match the sum of the extracted surface and interface sheet densities within error, so a device designer should treat these two regions as the main charge reservoirs.","A 5 nm In$_{0.84}$Al$_{0.16}$As cap removes about $2.2\\times10^{12}$ cm$^{-2}$ of surface charge, which means surface accumulation can be suppressed independently of the interface channel.","The interface sheet density of $4.8\\times10^{12}$ cm$^{-2}$ corresponds to a volume concentration about ten times lower than the comparable InAs-on-GaAs value, consistent with a lower dislocation density on the metamorphic buffer.","Strain relaxation and XRD linewidth both scale roughly as $1/t$, so the structural quality of these films is set by thickness broadening and relaxation, not by a thickness-independent defect density.","The model intentionally fails for 12.5 nm and 25 nm films because near the critical thickness fewer dislocations have nucleated, so the interface channel is less charged than the constant-$t_{\\mathrm{int}}$ picture assumes."],"supporting_citations":[{"why":"Supplies the two-layer Hall coefficient formula (Eq. 6) on which the parallel-channel model is based.","marker":"[28]"},{"why":"Provides the same functional dependence for two parallel channels and the interface-layer thickness and carrier-density reference values compared in the fit.","marker":"[27]"},{"why":"Source of the three-region picture (surface, bulk, interface) applied to InAs-type films.","marker":"[26]"},{"why":"Establishes that thickness-dependent Hall data reflect non-uniform carrier concentration along the growth direction, motivating the model.","marker":"[25]"},{"why":"Geometrical model used to interpret strain relaxation scaling as 1/t in the XRD data.","marker":"[22]"},{"why":"Growth procedure and critical-thickness estimate for InAs on the same metamorphic buffer family.","marker":"[17]"},{"why":"Provides transport coefficients and parameter values used in checking the factor f in Eq. (7).","marker":"[30]"}],"fun_headline_variants":["Thin InAs: most charge sits at surface and interface","Surface and interface hold most charge in thin InAs","Hall data shows InAs charge at interfaces, not bulk","Two-channel model: thin InAs carriers live at edges","InAs films: surface and interface dominate thin films"],"cache_read_input_tokens":3200,"weakest_assumption_plain":"The model assumes that the interface layer keeps a constant thickness of about 10 nm and that the carrier concentration of the combined bulk-surface layer is the same for every film thickness; the paper itself notes that the model overestimates the charge for the 12.5 nm and 25 nm samples, where the interface layer should be less charged near the critical thickness.","fun_headline_variants_meta":{"raw":{"variants":["Thin InAs: most charge sits at surface and interface","Surface and interface hold most charge in thin InAs","Hall data shows InAs charge at interfaces, not bulk","Two-channel model: thin InAs carriers live at edges","InAs films: surface and interface dominate thin films"]},"model":"deepseek-v4-flash","effort":"low","cost_usd":0.000317,"raw_usage":{"total_tokens":1853,"prompt_tokens":1065,"completion_tokens":788,"prompt_tokens_details":{"cached_tokens":384},"prompt_cache_hit_tokens":384,"prompt_cache_miss_tokens":681,"completion_tokens_details":{"reasoning_tokens":708}},"tokens_in":681,"tokens_out":788,"duration_ms":7386,"temperature":1.0,"reasoning_tokens":708,"cache_read_input_tokens":384,"cache_creation_input_tokens":0},"cache_creation_input_tokens":0},"created_at":"2026-08-10T15:45:06.222564+00:00","model_set":{"reader":"deepseek-v4-flash"},"falsifier":"Perform low-temperature Hall and mobility-spectrum measurements on the 50 nm and 100 nm samples to resolve the two channels independently; if the resolved surface and interface sheet densities do not sum to the measured total sheet density, or if $n_{\\mathrm{bs}}$ extracted from films of different thickness is found to change systematically as strain relaxes, the fixed-$n_{\\mathrm{bs}}$ two-channel decomposition in Eq. (8) is not a unique description.","supporting_citations":[{"cited_title":"Geometrical theory of critical thickness and relaxation in strained-layer growth","cited_arxiv_id":null,"evidence_quote":"Geometrical model used to interpret strain relaxation scaling as 1/t in the XRD data."},{"cited_title":"Theory of an experiment for measuring the mobility and density of carriers in the space-charge region of a semiconductor surface","cited_arxiv_id":null,"evidence_quote":"Supplies the two-layer Hall coefficient formula (Eq. 6) on which the parallel-channel model is based."},{"cited_title":"Thickness depend- ence of the structural and electrical properties of InAs layers epitaxially grown by MBE on GaAs (001)","cited_arxiv_id":null,"evidence_quote":"Provides the same functional dependence for two parallel channels and the interface-layer thickness and carrier-density reference values compared in the fit."},{"cited_title":"Transport properties of InSb and InAs thin films on GaAs substrates","cited_arxiv_id":null,"evidence_quote":"Source of the three-region picture (surface, bulk, interface) applied to InAs-type films."},{"cited_title":"Temperature- dependent transport properties of InAs films grown on lattice-mismatched GaP","cited_arxiv_id":null,"evidence_quote":"Establishes that thickness-dependent Hall data reflect non-uniform carrier concentration along the growth direction, motivating the model."},{"cited_title":"Transport coefficients of InAs epilayers","cited_arxiv_id":null,"evidence_quote":"Provides transport coefficients and parameter values used in checking the factor f in Eq. (7)."}],"review_version":1}